Method validation - Characterisation of phenolics in the Corylus extracts by HPLC-MS

5. R ESULTS

5.3. Characterisation of phenolics in the Corylus extracts by HPLC-MS

5.4.1. Method validation

The quantitative methods provided linear responses for all standards within the investigated range (Table 8). Retention time repeatability was suitable for all compounds, relative standard deviation was <0.15% (n = 6). Specificity of the methods was checked by injecting pure solvent. No co-elution was observed at the retention time of the analytes. Precision of the methods was tested by performing intra- and inter-day evaluation of solutions containing the target analyte in three concentrations (low, mid and high values of the calibration range), precision and accuracy tests were performed

in triplicate. The relative standard deviation for intra- and inter-day precision was <10%

for all quantitative methods, while intra- and inter-day accuracy ranged from 91.2% to 108.3% (Table 9).

Table 8. Regression, LOQ and LOD of the MRM quantitative methods (standard solutions were prepared in triplicates and injected once)

Standard Regression equation r² Regression

range (µg/ml) LOD

(µg/ml) LOQ (µg/ml)

Myricetin-3-O-rhamnoside y=1330.4x+1062.5 0.9999 0.1-300 0.020 0.067

Quercetin-3-O-rhamnoside y=1188.6x+2036 0.9996 0.1-300 0.010 0.033 Oregonin y = 2293.9x + 2405 0.9999 0.1-100 0.005 0.017 Hirsutenone y=332.82x+13.11 0.9977 0.01-50 0.002 0.007

Table 9. Method validation: Precision and accuracy of the quantitative methods Nominal conc. (µg/ml) Precision (RSD%) Accuracy (%)

70 5.4.2. Quantitative results

Quantity of myricetin-3-O-rhamnoside, quercetin-3-O-rhamnoside, hirsutenone and oregonin were determined by HPLC-MS/MS experiments using MRM (multiple reaction monitoring). Linear regression analyses were performed by using external calibration (Table 8). The quantifier and qualifier ions were designated by the evaluation of the mass spectra of the compounds. The qualifier/quantifier ion ratio remained within the ± 20% range of the determined value for each analyte (see sections 4.7.1. and 4.7.2.). Results of the quantitative analyses are presented in Table 10 as mean and standard deviation of three parallel measurements.

Table 10. Results of the quantitative analyses

Quantity in the extracts (mean±SD, n=3) (µg/mg extract)

Myricetin-3-O-rhamnoside

Quercetin-3-O-rhamnoside Oregonin Hirsutenone C. avellana leaves EtOAc 37.70±0.97 34.94±0.87 n.d. 2.08±0.03 C. avellana leaves MeOH 72.64±0.13 16.94±0.26 n.d. 0.33±0.01 C. avellana bark EtOAc 13.06±0.81 28.30±0.85 n.d. n.d.

C. avellana bark MeOH 2.42±0.08 15.52±0.74 n.d. n.d.

C. colurna leaves EtOAc 16.80±0.22 89.65±0.56 n.d. n.q.

C. colurna leaves MeOH 5.89±0.09 41.50±0.38 n.d. n.d.

C. colurna bark EtOAc 3.89±0.26 39.20±0.34 3.1±0.05 n.d.

C. colurna bark MeOH 1.65±0.19 7.86±0.11 2.1±0.03 n.d.

C. maxima leaves EtOAc 16.90±0.09 8.20±0.02 3.40±0.04 8.39±0.01 C. maxima leaves MeOH 30.10±0.80 19.80±0.75 2.21±0.21 0.59±0.05 C. maxima bark EtOAc 15.86±0.22 38.52±0.02 1.34±0.02 0.54±0.00 C. maxima bark MeOH 15.50±0.02 4.90±0.20 0.31±0.02 0.10±0.00 n.d.=not detected; n.q.=not quantified

5.5. HPLC-based DPPH scavenging assay

After spiking the Corylus samples with the DPPH radical solution (see section 4.8.), the decrease in the peak areas of the main compounds was examined. The chromatograms of C. colurna bark ethyl acetate extract are presented in Figures 21 and 22 (the latter presents the magnification of the original chromatograms for better recognition of the minor constituents) as an example.

The results of the HPLC based DPPH scavenger activity analyses are presented in Tables A2-13 in section 13.2. ΔArea was calculated according to eq. 1 and 2; and the ΔArea ratio according to eq 3.

ΔArea = peak area1-peak area2 (eq. 1) Decrease in the peak area (%) = (ΔArea/ peak area1)100% (eq. 2) ΔArea ratio= (ΔArea/Σ ΔArea)100% (eq. 3)

where ΔArea (mAus) is the change in the peak area; peak area1 (mAus) is the area of the compound’s chromatographic peak in the control sample; peak area2 (mAus) is the area of the compound’s chromatographic peak in the DPPH spiked sample.

Figure 21. HPLC-UV chromatograms of the control sample and the sample after spiking with DPPH of C. colurna bark ethyl acetate extract (CCBE) (see section 4.8.)

Figure 22. Enlarged chromatograms of the control sample and the sample after spiking with DPPH of C. colurna bark ethyl acetate extract (CCBE) (see section 4.8.) Compound 23: quercetin-3-O-rhamnoside, 27: kaempferol-3-O-rhamnoside, 31:

caffeoyl- hexoside, 36: catcechin/epicatechin, 37: procyanidin dimer

6. D

ISCUSSION

6.1. Quantitative phytochemical analyses

Based on the results of the quantitative determination of phenolics in the Corylus crude drugs (Figure 23, Table 10) by spectrometric methods described in Ph. Hg. VIII. (2003), it could be concluded that all the crude drugs contained notable amounts of polyphenol compounds, with C. maxima bark being the richest in these constituents (3.63±0.27g total phenolics in 100g crude drug). C. avellana leaves and bark were found to contain high quantities of both tannins and flavonoid derivatives. In the leaves tannins (1.37±0.05 g/100g crude drug), while in the bark flavonoids (1.54±0.05 g/100g crude drug) were presented in higher amounts. In the case of C. colurna samples, the opposite was observed; the leaves were found to be richer in flavonoids (0.49±0.13 g/100g crude drug), than in tannins (0.38±0.01 g/100g crude drug); while in the bark tannins (1.33±0.37 g/100g crude drug) were presented in higher quantity compared with flavonoids (0.43±0.05 g/100g crude drug). In C. maxima leaves the lowest amount of tannins was measured among all the crude drugs, both the leaves and bark were found to be richer in flavonoids (0.96±0.13 g/100g crude drug in the leaves and 0.82±0.09 g/100g crude drug in the bark), than in tannin derivatives (0.10±0.01 g/100g crude drug in the leaves and 0.51±0.06 g/100g crude drug in the bark).

It has to be noted that the plant samples of the three species were obtained from different places in Hungary and also at different times of the year; thus considering that both these factors have an influence on the quantity of phenolic compounds in the crude drugs, some caution must be taken when comparing data regarding different species.

Figure 23. Quantitative determination of phenolics in the Corylus crude drugs

6.2. Antioxidant activity assays

The antioxidant activity of plant phenolics is widely reported, besides, the extracts of C.

avellana and a natural mixture of flavonoids isolated from C. colurna have been shown to posses potent antioxidant activity (Delgado et al. 2010, Benov and Georgiev 1994).

Several mechanisms have been proved to play role in the antioxidant effect of polyphenols that include free radical scavenging (ROS or RNS); suppressing ROS/RNS formation by enzyme inhibition or by chelating of trace elements; and also the enhancing of the antioxidant defence system (Cotelle 2001). The first effect mentioned above, the deactivation of free radicals, can occur by two main mechanisms, hydrogen atom transfer (HAT) and single electron transfer (SET) reactions. The dominant mechanism is determined by the structure and solubility of the antioxidant as well as the method of testing (Prior et al. 2005). The efficacy of the antioxidants is mainly determined by the bond dissociation energy (BDE) of their reactive functional group and the ionisation potential (IP).

Since the interest in antioxidants is progressively increasing, several methods have been developed and improved for the determination of free radical scavenging activity (Magalhaes et al. 2008, Prior et al. 2005). The HAT-based methods measure the ability of the antioxidant to quench free radicals by the donation of hydrogen. The reactivity

present in these assays mainly depends on the BDE of the H-donating group in the molecules (Wright et al. 2001, Huang et al 2005). The SET-based methods determine the ability of an antioxidant to transfer an electron to reduce radicals. This latter mainly depends on the IP of the antioxidant investigated. The in vitro tests using DPPH and ABTS as free radicals were chosen for the determination of the antioxidant activity of the Corylus extracts prepared with ethyl acetate and methanol. Although the DPPH and ABTS assays are usually classified as SET reactions these two radicals can also be neutralised by H atom transfer, with the latter occurring as a marginal reaction pathway (Ou et al 2005). Although the DPPH assay has been developed in 1958 (Blois 1958), it still has widespread application (Kedare et al. 2011, Thaipong et al. 2006, Brand-Williams et al. 1995). The main withdrawal of the method is that steric accessibility is considered one of the major determinants of the reaction, meaning that smaller molecules have better access to the radical site, thus show better scavenger capacity in the test (Prior et al. 2005). The latter method, that utilises ABTS as stable free radical, was found to be inappropriate for the quantitative determination of the antioxidant capacity, although it was successfully utilised to provide a ranking order of different antioxidants (Miller et al. 1996). The main advantages of the two methods mentioned above include rapid reaction with the antioxidants (1-30 min) (Brand-Williams et al.

1995, Re et al. 1999); the fact that both work well in a methanol or ethanol solvent system, which is important considering the solubility of the phenolics examined in our study; furthermore both tests are operationally simple (Awika et al. 2003). Since our aim was the preliminary investigation of the scavenger capacity of the Corylus extracts as well as the comparison of their activity, both the tests were found to be suitable for our experiments.

Based on the results (Tables 3 and 4) it could be concluded that all the extracts possessed notable activity in both in vitro tests (with IC50 values not higher than 50 μg/ml). For reasons of clarity, the antiradical power (1/IC50, ml/µg) of the extracts is depicted in Figure 24: the larger the antiradical power, the more efficient the antioxidant.

Figure 24. Antiradical power (1/IC50, ml/µg) of the Corylus extracts in the DPPH and ABTS in vitro tests

CALE: C. avellana leaves ethyl acetate extract, CALM: C. avellana leaves methanolic extract, CABE: C.

avellana bark ethyl acetate extract, CABM: C. avellana bark methanolic extract; CCLE: C. colurna leaves ethyl acetate extract, CCLM: C. colurna leaves methanolic extract, CCBE: C. colurna bark ethyl acetate extract, CCBM: C. colurna bark methanolic extract; CMLE: C. maxima leaves ethyl acetate extract, CMLM: C. maxima leaves methanolic extract, CMBE: C. maxima bark ethyl acetate extract, CM BM: C. maxima bark methanolic extract.

The methanolic extracts of C. maxima leaves and bark showed significantly the highest scavenger capacity among the Corylus samples in the ABTS test. The bark samples acted noticeably stronger against ABTS free radical than the leaves, while the difference in the DPPH test was not significant. Both the leaves and bark methanolic extracts showed higher antioxidant activity than the ethyl acetate extracts in both in vitro tests.

The C. colurna bark extracts showed the highest scavenging capacity among the Corylus samples in the DPPH test, exceeding also the activity of all the standards except hirsutenone. The scavenger capacity of the bark extracts was found to be higher than that of the leaves in the ABTS test as well. The ethyl acetate extract of the leaves possessed more potent scavenger activity against DPPH free radical, while the methanolic extract acted stronger against ABTS. In case of the bark samples the difference between the activities of the ethyl acetate and methanolic extract was not explicit.

In case of C. avellana significant difference was not observed between the DPPH scavenging capacities of the extracts, while in the ABTS tests the bark samples showed

noticeably higher antioxidant activity than the leaves. Although neither in the latter case was significant difference found between the activities shown by the ethyl acetate and methanolic extracts of the bark or leaves.

It has to be noted that neither the DPPH nor the ABTS test can be considered physiological, since none of the two radicals occur in biological systems. Therefore, the results obtained by these methods cannot be extrapolated to in vivo data. However, the measured high in vitro scavenging capacity regarding the extracts indicated the presence of potential natural antioxidants, furthermore the fact that the measured antiradical activities did not show any trend, indicated notable differences in the phenolic profile of the extracts. Thus, characterisation of the phenolic fingerprint of the samples was found to be reasonable.

6.3. Characterisation of phenolics in the Corylus extracts by HPLC-MS

Discussion of the characterisation of phenolics detected in the Corylus extracts is based on structural groups not on plant species. Characterisation was based on results presented in Tables 5-7 (see section 5.3.).

6.3.1. Diarylheptanoids

UV spectra with absorption maxima at 250-260 and 300-310 nm together with characteristic mass spectra indicated diarylheptanoid structures in the case of twenty compounds of the Corylus extracts. No previous literature data was found regarding diarylheptanoids in any of the investigated Corylus species, thus all the compounds described below are reported in these plants for the first time. From the accurate molecular mass and formula given by ESI-TOF, and fragmentation patterns acquired by collision-induced dissociation (CID) in ESI-MS/MS analyses compared to authentic standards and to literature data (Jiang et al. 2006) we were able to characterise the structures of fifteen diarylheptanoids, although the applied MS/MS method is not suitable for the accurate identification of the molecules where no matching standards were available.

Figure 25. Assumed (-) ESI-MS/MS fragmentation of diarylheptanoids detected in the Corylus extracts.

Although the UV spectra and the molecular formulas of compounds 16-20 indicated diarylheptanoid aglycones, from the fragmentation pattern given by ESI-MS/MS analyses no appropriate conclusions could have been drawn about their structures.

Among the fifteen compounds hirsutenone and oregonin were identified by comparing their chromatographic and spectrometric data to authentic standards. Since for other diarylheptanoid compounds no matching standards were available, tentative identification was based on comparison of their mass spectral data to hirsutenone and oregonin and other diarylheptanoids being present in the extracts. In general two fragmentation pathways (pathways A and B, see Fig. 25) were observed regarding these

compounds. In the case of compounds 1-12 the same trends in the CID were observed (A), while compounds 13-15, where the presence of an unsubstituted hydroxyl and a keto group on the alkyl chains at positions 5 and 3, respectively, was assumed showed different fragmentation behaviour (B).

The molecular ion [M-H]^- of hirsutenone (1) was detected at m/z 327.1228, the characteristic product ions at m/z 205 and 179. The ESI-TOF conjunction analysis and the molecular formula calculation pointed to the formula C19H20O5. The (-) ESI-MS/MS fragmentation shown in Figure 26 was supposed based on literature data (Jiang et al.

2006). Identification of hirsutenone in the Corylus extracts was performed by comparison of chromatographic and mass spectrometric behaviour to those of an authentic standard, and also by spiking the sample solutions with the standard in two different chromatographic methods (see sections 4.6.1. and 4.6.2.).

Figure 26. (-) ESI-MS/MS fragmentation of hirsutenone (1).

The molecular ion [M-H]^- of oregonin (2) was detected at m/z 477.1779, exhibited characteristic fragment ions at m/z 327, 205 and 179. The ESI-TOF conjunction analysis and the molecular formula calculation pointed to the formula C24H30O10. The neutral loss of 150 amu between m/z 477 and m/z 327 referred to a pentose sugar residue, while the aglycone fragment ion at m/z 327 and the characteristic fragment ions at m/z 205 and 179 indicated hirsutenone aglycone (Fig. 27). Identification of oregonin in the Corylus extracts was based on comparison of chromatographic and mass spectrometric behaviour to those of an authentic standard, and was also performed by spiking the sample solutions with the standard in two different chromatographic methods (see sections 4.6.1. and 4.6.2.).

Figure 27. (-) ESI-MS/MS fragmentation of oregonin (2).

Compound 3 exhibited molecular ion at m/z 507.1855 and characteristic fragment ions at m/z 327 and 205. The ESI-TOF conjunction analysis and the molecular formula calculation pointed to the formula C25H32O11. The neutral loss of 180 amu indicated a hexose moiety, while the aglycone fragment at m/z 327 and the characteristic fragment ion at m/z 205 indicated hirsutenone aglycone. Therefore, similar fragmentation (Fig.

28) was assumed as in the case of oregonin and compound 3 was tentatively identified as hirsutanolol-hexoside.

Figure 28. Assumed (-) ESI-MS/MS fragmentation of hirsutanolol-hexoside (3).

Compounds 4 and 5 exhibited molecular ions [M-H]^- at m/z 461.1788 and characteristic fragment ions at m/z 311, 205 and 189. The ESI-TOF conjunction analysis and the molecular formula calculation pointed to the formula C24H30O9 for both compounds, which differs from the molecular formula of oregonin with one oxygen atom. The neutral loss of 150 amu refers to a pentose sugar residue. In the aglycone fragment ions (m/z 311) a mass shift of 16 Da with reference to the molecular ion of hirsutenone (m/z 327) was observed, suggesting the absence of one oxygen atom in the aglycone molecules compared with hirsutenone. Based on the detection of the two other abundant product ions (m/z 205 and 189), the presence of only one hydroxyl group on one of the benzyl rings of the compounds and also the presence of a mixture of two structural isomers was supposed (Fig. 29). Compound 4 was tentatively identified as 1-(4-hydroxyphenyl)-7-(3,4-dihydroxyphenyl)-heptan-3-one-5-O-pentoside and

compound 5 as 1-(3,4-dihydroxyphenyl)-7-(4-hydroxyphenyl)-heptan-3-one-5-O-pentoside.

Figure 29. Assumed (-) ESI-MS/MS fragmentation of 1-(4-hydroxyphenyl)-7-(3,4-dihydroxyphenyl)-heptan-3-one-5-O-pentoside (4)

and 1-(3,4-dihydroxyphenyl)-7-(4-hydroxyphenyl)- heptan-3-one-5-O-pentoside (5).

UV spectrum of compound 6 pointed to diarylheptanoid structure. It exhibited molecular ion [M-H]^- at m/z 473.1797 and an aglycone fragment ion at m/z 293. The ESI-TOF conjunction analysis and the molecular formula calculation corresponded to the formula C25H30O9. The neutral loss of 180 amu indicated a hexose moiety. Based on the (-) ESI-MS/MS spectrum, the compound was characterised as 5-hydroxy-1,7-bis-(4-hydroxyphenyl)-6-hepten-3-one-hexoside, which was also detected in aglycone form (12).

The molecular ion [M-H]^- and characteristic product ions of platyphyllonol-pentoside (7) were detected at m/z 445.1868 and at m/z 295 and 189, respectively. The ESI-TOF conjunction analysis and the molecular formula calculation pointed to the formula C24H30O8 which differed from the structure of oregonin (C24H30O10) with two oxygen atoms. The neutral loss of 150 amu indicated a pentose sugar residue. In the aglycone fragment ion (m/z 295) a mass shift of 32 Da was observed with reference to the molecular ion of hirsutenone (m/z 327), while in the other characteristic product ion (m/z 189) a mass shift of 16 Da was detected with reference to the most intense product ion of hirsutenone (m/z 205). These observations suggested the presence of only one hydroxyl group on each benzyl rings (in the case of hirsutenone two hydroxyl groups are presented on both benzyl rings). Diarylheptanoids platyphyllon and platyphyllonol-glycosides had already been reported from Betulaceae plants (Novaković 2013),

therefore, the presence of the hydroxyl groups was supposed in position 4 of the benzyl rings and compound 7 was characterised as platyphyllonol-pentoside (Fig. 30).

The molecular ion [M-H]^- and characteristic product ions of platyphyllonol-hexoside (8) were detected at m/z 475.1974 and at m/z 295 and 189, respectively. The ESI-TOF conjunction analysis and the molecular formula calculation pointed to the formula C25H32O9. The neutral loss of 180 amu indicated a hexose sugar residue, while the aglycone fragment ion and the other characteristic product ions indicated platyphyllonol aglycone (Fig.30.).

Figure 30. Assumed (-) ESI-MS/MS fragmentations of platyphyllonol-pentoside (7) and platyphyllonol- hexoside (8).

Compound 9 exhibited molecular ion [M-H]^-at m/z 331.1545, while compound 10 at m/z 329.1383. In the molecular ion of compound 9 a shift of 4 Da was observed with reference to the molecular ion of hirsutenone (m/z 327), while the calculated molecular formulas (C19H24O5 and C19H20O5, respectively) showed a difference of 4 hydrogen atoms. Based on the (-) ESI-MS/MS spectra fragmentation shown in Figure 31 was assumed and compound 9 was tentatively identified as 3-hydroxy-1,7-bis-(3,4-dihydroxyphenyl)-heptan. Both the exact masses and the molecular formulas of compounds 9 and 10 (C19H24O5 and C19H22O5, respectively) suggested the difference of two hydrogen atoms between the two compounds and possibly a presence of a double bond in the alkyl chain in case of compound 10. In the molecular ion of compound 10 (m/z 329) a shift of 2 Da was observed with reference to the molecular ion of hirsutenone (m/z 327), while in the other characteristic product ion (m/z 207) also a mass shift of 2 Da was detected with reference to the most intense product ion of hirsutenone (m/z 205). Based on this data the presence of a hydroxyl group in position 5 and a double bond in the alkyl chain was supposed and compound 10 was tentatively

identified as 3-hydroxy-1,7-bis-(3,4-dihydroxyphenyl)-hepten (Fig. 32.). Based on the molecular formulas and the (-) ESI-MS/MS fragmentation the possible structure 1,7-bis-(3,4-dihydroxyphenyl)-hepta-3-one for compound 10 cannot be excluded, but in that case the co-elution with compound 9 in the applied chromatographic system would be rather unlikely.

Figure 31. Assumed (-) ESI-MS/MS fragmentation of 3-hydroxy-1,7-bis-(3,4-dihydroxyphenyl)-heptan (9).

Figure 32. Assumed (-) ESI-MS/MS fragmentation of 3-hydroxy-1,7-bis-(3,4-dihydroxyphenyl)-hepten (10).

The molecular ion [M-H]^- and characteristic fragment ions of compound 11 were detected at m/z 311.1283, and at m/z 205 and 189, respectively. The ESI-TOF conjunction analysis and the molecular formula calculation correspond to the formula C19H20O4, which differed from that of hirsutenone (C19H20O5) by one oxygen atom. The (-) ESI-MS/MS spectrum of the compound suggested the presence of only one hydroxyl group on one of the benzyl rings. Product ions at m/z 205 and 189 were produced by different neutral moiety losses, because either ring could be deprotonated during ionisation and hold the negative charge (Fig. 33.). Although it has to be noted that according to the (-) ESI-MS/MS spectrum the co-elution of two structural isomers 1-(4-hydroxyphenyl)-7-(3,4-dihydroxyphenyl)-hept-4-en-3-one (11a) and 1-(3,4-dihydroxyphenyl)-7-(4-hidroxyphenyl)-hept-4-en-3-one (11b) is also possible.

Figure 33. Assumed (-) ESI-MS/MS fragmentation of 1-(3,4-dihydroxyphenyl)-7-(4-hydroxyphenyl)- hept-4-en-3-one (11).

The deprotonated molecular ion [M-H]^- of hirsutanolol (13) was detected at m/z 345.1327. The ESI-TOF conjunction analysis and the molecular formula calculation pointed to the formula C19H22O6, which differed from the formula of hirsutenone with two hydrogen and one oxygen atom. The characteristic fragment ions at m/z 205, 179 and 165 were also detected in the (-) ESI-MS/MS spectrum of the compound. The presence of a hydroxyl and a keto group on the alkyl chain at positions 5 and 3,

In document Recovery of new diarylheptanoid sources in Betulaceae Characterisation of the phenolic profile of Corylus species by HPLC-ESI-MS methods (Pldal 68-0)